Kenneth Stanley Uninterrupted: Why Greatness Cannot Be Planned: The Myth of the Objective

  Рет қаралды 3,650

Clint Chilcott

Clint Chilcott

8 жыл бұрын

7 minutes of interruptions removed!
Buy his book!

Пікірлер: 8
@anjankatta1864
@anjankatta1864 3 жыл бұрын
Thank the lord for making this, makes it so much easier to share
@thunksalot
@thunksalot 2 жыл бұрын
Yes. Thank you. That moderator was TERRIBLE. Never seen someone interrupt a presenter so much and so needlessly.
@Karoltos
@Karoltos Ай бұрын
Do you have the full 40-minute video somewhere? The original video is no longer available, and I’m actually looking for Kenneth’s response to one of the questions from audience :D
@DubFunTheatre
@DubFunTheatre 6 жыл бұрын
Why can't it be that we just don't know what are the important characteristics to look for? They were looking for "something that looks must like a butterfly" from the very beginning, but watch any artist paint a painting and it's not as if their painting looks like some form of their end result the entire way. If I was looking for a skull, the first few steps would be to increase the line density of the painting first, then it might be any number of "success" criteria, each with branching future success criteria. For example, If I found one success with a "find a generally round swiggly line", among the next success criteria options would be "hollow out the center, but preferably with two distinct dots".
@bokaratom
@bokaratom 2 жыл бұрын
Great stuff! I have been searching for a mechanism to make the Evolutionary Art to not be supervised learning. I am tired of going through literally 10's of thousands of images saying I like this and I don't like this. I now need to find some way to score them on how different they are from the rest of the population and use that as the fitness function.
@bokaratom
@bokaratom 2 жыл бұрын
It will probably create a population that I can use as a stepping stone and just finish off with the I like this and don't like that approach. It's interesting that the evolution of art seems to have worked better than the evolution of trading systems. I always thought it was cheating or even a cop out because anything can be called good art (it's subjective) but not everything can be called a good trading system (it's less subjective).
@bokaratom
@bokaratom 2 жыл бұрын
In reality it may be that the advantage of the evolution of art was that it was not driven by a goal. There was less of a concept of what the outcome should be.
@bokaratom
@bokaratom 2 жыл бұрын
That being said I will mention that in the video Ken is disdainful of the art created by consensus as compared to the art created by picbreeder. I preferred the consensus art. Ken is excited by pictures that represent objects with which he's familiar ,i.e., cars teapots butterflies etc. I often prefer abstract representations swirls of color and form that invoke emotion or feelings of beauty or calm as apposed to familiarity with known objects. 🤠
#038 - Prof. KENNETH STANLEY - Why Greatness Cannot Be Planned
2:46:27
Machine Learning Street Talk
Рет қаралды 140 М.
одни дома // EVA mash @TweetvilleCartoon
01:00
EVA mash
Рет қаралды 5 МЛН
Something Strange Happens When You Follow Einstein's Math
37:03
Veritasium
Рет қаралды 7 МЛН
What is ChatGPT doing...and why does it work?
3:15:38
Wolfram
Рет қаралды 2,1 МЛН
Dr. Adam Grant: How to Unlock Your Potential, Motivation & Unique Abilities
3:12:22
1. Introduction to 'The Society of Mind'
2:05:54
MIT OpenCourseWare
Рет қаралды 1,4 МЛН
Deep Learning: A Crash Course (2018) | SIGGRAPH Courses
3:33:03
ACMSIGGRAPH
Рет қаралды 2,6 МЛН
🕳️ [Christophe Galfard] Des trous noirs à l'origine de l’Univers
2:16:20
Espace des sciences
Рет қаралды 6 МЛН
Machine Learning for Everybody - Full Course
3:53:53
freeCodeCamp.org
Рет қаралды 5 МЛН
All New Atlas | Boston Dynamics
0:40
Boston Dynamics
Рет қаралды 5 МЛН
Vortex Cannon vs Drone
20:44
Mark Rober
Рет қаралды 13 МЛН